Today’s Topics

  1. Customizing visualizations
  2. Making static visualizations interactive with plotly
  3. Creating interactive maps with leaflet

Setup

Loading in our Libraries

knitr::opts_chunk$set(echo = TRUE)
library(tidyverse)
library(here)
library(janitor)
library(plotly)
library(rnaturalearth)
library(leaflet)
library(sf)
library(vembedr)

Read in 2023 Sustainable Development Data with read_csv() and here()

sdr_data <- read_csv(here("data/SDR-2023-Data.csv"))

Clean column names

sdr_data <- sdr_data %>% 
  clean_names()

Customize a Histogram

Let’s start with the histogram you created last lesson with ggplot() and geom_histogram()

ggplot(sdr_data, aes(x = goal_4_score, fill=regions_used_for_the_sdr)) +
  geom_histogram()

Looks Nice, but let’s improve it!

  • add a theme with theme_minimal()
  • add a color palette with scale_fill_viridis_d()
  • add a title, x axis label, y axis label, and legend label with labs()

We start with the same exact code and add to it with +

ggplot(sdr_data, aes(x = goal_4_score, fill=regions_used_for_the_sdr)) +
  geom_histogram() +
  theme_minimal() +
  scale_fill_viridis_d() +
  labs(title = "Distributions of SDG 4 Scores",
       x = "SDG 4 Score",
       y = "Number of Countries",
       fill = "Region")

Awesome, looking much better!

  • However, it’s still a little tricky to tell exactly how many countries in each region have each SDG 4 Score
  • There’s a few ways to go address this, one of them is to make the plot interactive with ggplotly()

Making static visualizations interactive

Interactive visualizations are a really exciting part of data science! Interactivity engages the viewer by allowing them to explore the data in a way they cannot with static visualizations

  • The great part is that with the ggplotly() function from the plotly package, making interactive visualizations is very simple

    • First we create the plot with the exact same code

      • The only difference is that we assign the plot a name with <- the same way we do with dataframes or lists

      • Next we put the name that we give the plot into the ggplotly function

goal_4_histogram <- ggplot(sdr_data, aes(x = goal_4_score, fill=regions_used_for_the_sdr)) +
  geom_histogram() +
  theme_minimal() +
  scale_fill_viridis_d() +
  labs(title = "Distributions of SDG 4 Scores",
       x = "SDG 4 Score",
       y = "Number of Countries",
       fill = "Region")

ggplotly(goal_4_histogram)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## Warning: Removed 27 rows containing non-finite values (`stat_bin()`).

Epic! Now we can hover over the histogram and get some info

  • For example: We can see 2 countries in the LAC region have a score of 68

We can also double-click the boxes in the legend to only view the regions we’re interested in

Challenge 1

Here is the code for a scatter plot you made in the last lesson:

ggplot(data = sdr_data, aes(x= goal_3_score, y = goal_4_score, color = regions_used_for_the_sdr)) +
  geom_point()

Edit the code chunk above that generates the plot to:

  • add a theme with theme_minimal()
  • add a color palette with scale_fill_viridis_d()
  • add the appropriate title, x axis label, y axis label, and legend label with labs()

AND

  • make it interactive with ggplotly()

    • hint: you will have to assign a name to the plot

Creating interactive maps with leaflet

This takes a few fun steps to do

  1. We need data on countries and their physical location on the globe

  • Lucky for us the rnaturalearth package has this information for us

    • We use the ne_countries() function to access the data and we assign it the name world
world <- ne_countries(scale = "medium", returnclass = "sf")

  • Now we have a dataframe named world in our environment

    • It has 241 locations/countries (rows) and 64 columns describing each location/country

    • We are only interested in 3 columns:

      • Country Name
      • ISO3 codes which are three-letter country codes
      • Geometry
        • This column contains the info describing a country’s physical location on the globe

Let’s select the three columns we are interested, making the dataframe much smaller

world <- world %>% 
  select(name_long, iso_a3, geometry)

We’re almost ready to join these dataframes

  • We have ISO3 codes in both of our dataframes (sdr_data and world) but the columns that contain the ISO3 codes have different names

    • In the world dataframe it is iso_a3
    • In the sdr_data dataframe it is country_code_iso3

Let’s change the column in sdr_data called country_code_iso3 to iso_a3 to match the world dataframe

# Rename a column in a data frame or matrix
colnames(sdr_data)[which(colnames(sdr_data) == "country_code_iso3")] <- "iso_a3"

Perfect! Now we can join

Let’s use the left_join() function

  • We use left join because we want the 3 columns we are interested in from world to be joined/added to sdr_data

    • There are 35 countries/places in the world dataframe that are not in our sdr_data. Using full_join() would include these countries/places in the new joined dataframe, however, all the SDG info would be NA. TO avoid this, we’ll use left_join()

    • We’ll name the new joined dataframe sdr_data_world_joined and we will join by the column the 2 dataframes share: iso_a3

sdr_data_world_joined <- left_join(sdr_data, world, by = "iso_a3")

Nice!

The next 4 code chunks are slightly technical regarding the class of sdr_data_world_joined. We can check the class with the ‘class()’ function

class(sdr_data_world_joined)
## [1] "spec_tbl_df" "tbl_df"      "tbl"         "data.frame"

Right now, sdr_data_world_joined is a dataframe. In order to use the leaflet package that generates the map, we need sdr_data_world_joined to be an sf dataframe. sf stands for spatial features. Converting sdr_data_world_joined to class sf dataframe is easy with the st_as_sf() function and it allows leaflet to interpret the geometry column as something it can map

sdr_data_world_joined <- st_as_sf(sdr_data_world_joined)

Now lets check the class again to make sure sdr_data_world_joined is of class sf dataframe

class(sdr_data_world_joined)
## [1] "sf"          "spec_tbl_df" "tbl_df"      "tbl"         "data.frame"

The last step before mapping with leaflet is specifying a coordinate reference system (crs). There are many different coordinate refernece systmes, and here we choose WGS84

  • WGS84 stands for World Geodetic System 1984. It’s a global standard for measuring locations on Earth’s surface. WGS84 defines a coordinate system using longitude and latitude coordinates, where longitude measures east-west position and latitude measures north-south position. This system is widely used in GPS (Global Positioning System) devices, mapping applications, and geographic information systems (GIS) because it provides a consistent and accurate way to represent locations anywhere on Earth.
sdr_data_world_joined <- st_transform(sdr_data_world_joined, "+proj=longlat +datum=WGS84")

Awesome! We’re ready to make a map

  • Let’s map SDG 7 Scores

    • the my_text part of the code determines what happens when we hover over a country

    • the leaflet() function generates the map from the sdr_data_world_joined dataframe

      • we specify SDG 7 scores in the “color =” argument of the function
mytext <- paste(
    "Country: ", sdr_data_world_joined$country,"<br/>", 
    "Goal 7 Score: ", round(sdr_data_world_joined$goal_7_score, 2), 
    sep="") %>%
  lapply(htmltools::HTML)

leaflet(sdr_data_world_joined) %>% 
  addTiles()  %>% 
  setView( lat=10, lng=0 , zoom=2) %>%
  addPolygons(stroke = FALSE, fillOpacity = 0.5, smoothFactor = 0.5, color = ~colorQuantile("YlOrRd", goal_7_score)(goal_7_score), label = mytext)

Challenge 2

Can you edit the code chunk above that generates the map to: